flow bench testing
05-24-2010, 01:04 PM
#46
Race Car
OK, now I feel like I was sleeping during class! Quoted like that I would have to disagree with myself. I should have put it better!
Its easy to get too fixated on one aspect of a test result. Its also easy to come up with general rules about numbers that are a little too general….kind of like that quote!
The same thing happens with dyno charts – people can get fixated with the peak HP and miss the big picture. So what’s a better way to look at the results? I would start with the ‘area under the line’ approach. To take it a step further you need to know the cam specs to figure out the time function of the flow. Combine those two and you could do a quick and dirty total cubic feet per valve event calculation. Once you do that then some of the advantages of particular flow rates will make more sense. That works great for the intake side which is, in theory, under a constant pressure differential. The exhaust has a different dynamic – the pressure differential goes from a very high peak and eventually equalizes (or close to equal). That calculation can get pretty complex with a lot of assumptions.
Looking at the intake flow on the 16v head you can see that the flow is valve opening limited up to .300” and then starts to become port limited (warning – this is very generalized!). The port flow hasn’t become ‘choked’ at max cam lift so there is some benefit available by adding more lift to the cam – but not a huge amount. If you could make up a cam that simply opened the valves quicker but kept that same max lift and duration you would see a lot more total flow per valve event. Then you have to figure out the complex valve train dynamics….thats a whole ‘nother discussion!
The summary is that low lift numbers are important because of the time frame in the valve event that they occur (67’s exact point!)
No bellmount? – the intake runner is bellmouthed into the plenum…and I thought about putting one on the end of the manifold – but a previous comparison test showed that it made no difference in the flow bench readings so I don’t bother with it anymore.
Thanks for the catch ’67, after seeing it quoted like that I would not want folks to go away with the wrong idea
Its easy to get too fixated on one aspect of a test result. Its also easy to come up with general rules about numbers that are a little too general….kind of like that quote!
The same thing happens with dyno charts – people can get fixated with the peak HP and miss the big picture. So what’s a better way to look at the results? I would start with the ‘area under the line’ approach. To take it a step further you need to know the cam specs to figure out the time function of the flow. Combine those two and you could do a quick and dirty total cubic feet per valve event calculation. Once you do that then some of the advantages of particular flow rates will make more sense. That works great for the intake side which is, in theory, under a constant pressure differential. The exhaust has a different dynamic – the pressure differential goes from a very high peak and eventually equalizes (or close to equal). That calculation can get pretty complex with a lot of assumptions.
Looking at the intake flow on the 16v head you can see that the flow is valve opening limited up to .300” and then starts to become port limited (warning – this is very generalized!). The port flow hasn’t become ‘choked’ at max cam lift so there is some benefit available by adding more lift to the cam – but not a huge amount. If you could make up a cam that simply opened the valves quicker but kept that same max lift and duration you would see a lot more total flow per valve event. Then you have to figure out the complex valve train dynamics….thats a whole ‘nother discussion!
The summary is that low lift numbers are important because of the time frame in the valve event that they occur (67’s exact point!)
No bellmount? – the intake runner is bellmouthed into the plenum…and I thought about putting one on the end of the manifold – but a previous comparison test showed that it made no difference in the flow bench readings so I don’t bother with it anymore.
Thanks for the catch ’67, after seeing it quoted like that I would not want folks to go away with the wrong idea
A few things you hit on - could not possibly agree more about the dyno thing. Last engine I did work on was for a guy most folks here know. It was a 3.0L BMW. He asked me to evaluate a cam that a German company was telling him to run. I didn't like it at all. When we looked at his torque curve, his trans ratios, and where he was spending time on the track, we did a new intake, instead. Yes, we picked up 50 horsepower (306 to 357 at wheels), but more importantly, he beat the class record at Road Atlanta the first time out - by two seconds. There is only one guy he races against who beats him, and he is two classes above him.
Anyway, the stuff you hit on about valve lift regions, and being 0.300" is hitting into what us engineering types refer to the three zones of valve lift. In the first zone, the area is determined by the seat width and angle. The limiting area is the perpindicular line between the seats on the valve and the head seat (or seat seat or insert seat - the part that is machined on the head). It creates a cone that has a 90 degree angle (at the point, well below the valve). As the valve opens more, that line no longer is perpindicular - it is the line between the inside of the valve seat, and the outside of the head seat. As the valve opens more and more, that cone's angle gets less and less, as it starts to work its way to creating a cylinder. It would never become one, but that point becomes irrelevant as the choke point in the port becomes smaller than the area of that cone. Interesting fact - the big domestic pushrod engines like GM's LS series and the Hemi's never become port limited. The valve is so big that the curtain area is ALWAYS the limiting factor, rather than the port.
Of course, all of this stuff is simple geometry - flow numbers (or even better - discharge coefficients, which are normalized to size) start measuring the quality of the port and valve job.
I've done a lot of modelling in Excel looking mostly at geometry and velocities. I've done some stuff adding discharge coefficients into the mix, but very few folks get multiple pressures, and the Flow2=Flow1(P2/P1)^0.5 rule of thumb doesn't really work so well for ports and valve area stuff.
